# Why can't I stop scrolling? What the research actually shows

> Your willpower plays a smaller role than you think, and the dopamine story is shakier than the internet claims.

Source: https://ivehadit.app/blog/why-cant-i-stop-scrolling-research · Habits · Published: 2026-10-26

If you've typed "why can't I stop scrolling" into a search bar at midnight, the short answer is that three things stack up. The feed has no natural end, and European regulators now say that design pushes people into "autopilot mode". Scrolling is also a habit, so it starts before you've decided anything, and it gets stronger the more you do it. On top of that, most people have a real gap between how much they want to scroll and how much they actually do.

So nothing is uniquely wrong with you. In the largest experiment on this question, self-control problems explained about a third of the average person's social media use ([Allcott, Gentzkow and Song](https://web.stanford.edu/~gentzkow/research/DigitalAddiction.pdf)). That is a real effect. It also means most of your scrolling is time you'd probably choose anyway, which is worth remembering before the self-blame kicks in.

**48 minutes a day**: Average social media use driven by self-control problems, about 31% of total use, in a 12-week experiment with roughly 2,000 US adults ([Allcott, Gentzkow and Song, American Economic Review (2022)](https://web.stanford.edu/~gentzkow/research/DigitalAddiction.pdf))

## Why can't I stop scrolling? Start with a feed that has no bottom

In February 2026 the European Commission made preliminary findings that TikTok's addictive design breaches the Digital Services Act ([European Commission](https://digital-strategy.ec.europa.eu/en/news/commission-preliminarily-finds-tiktoks-addictive-design-breach-digital-services-act)). It named four features: infinite scroll, autoplay, push notifications and a highly personalised recommender. In July it said the same about Instagram and Facebook, naming the same four ([European Commission](https://digital-strategy.ec.europa.eu/en/news/commission-preliminarily-finds-addictive-design-instagram-and-facebook-breach-digital-services-act)).

The Commission's explanation of why scrolling is so addictive is plainer than most blog posts. By constantly "rewarding" users with new content, it says, these features fuel the urge to keep going and shift people into "autopilot mode". A book has chapters and a TV episode has credits. A feed only ends when something outside it interrupts you.

One caveat: "preliminary" means these are findings the companies can still respond to, with no final ruling yet. Even so, a regulator looked hard at the design and singled out the endless part. That should tell you something about whose fault the lost evening is.

## Scrolling is a habit, and the habit grows with use

The Allcott experiment followed about 2,000 adults with Android phones for roughly 12 weeks in 2020, tracking Facebook, Instagram, Twitter/X, Snapchat, YouTube and browsers. Some people were paid to cut back. While the payments ran, their use dropped by 56 minutes a day, or 39%.

The interesting part came after the money stopped. Use stayed 19 minutes a day lower over the next three weeks, and 12 minutes lower in the three weeks after that. The authors read this as evidence of habit formation: what you did yesterday shapes what you do today.

**19 minutes a day**: How much lower people's use stayed in the three weeks after payments to cut back had ended ([Allcott, Gentzkow and Song, American Economic Review (2022)](https://web.stanford.edu/~gentzkow/research/DigitalAddiction.pdf))

This usually gets told as bad news, and it is half of it. The same mechanism that makes tonight's scroll likelier after last night's also makes tomorrow easier if you cut back today.

## You scroll more than you mean to, and you guess low

That 48-minute average hides a wide spread. For 22% of participants, the excess was under 10 minutes a day. For 32%, it was more than an hour. If you're in the second group, you're far from alone, and if you're in the first, your scrolling may bother you less than the headlines suggest.

People also misjudge themselves. The control group underestimated its own next-period use by 6.1 minutes a day (about 4%), even after being reminded of recent use. Small, but it always points the same way: we expect future us to scroll less.

The study has limits. Participants were recruited through Facebook and Instagram ads, all used Android, and it ran during the pandemic. The authors report that reweighting the sample to look like the US population makes the estimated effect larger, and that effects stayed stable over time. The [Hamilton Project](https://www.hamiltonproject.org/publication/paper/digital-addiction-evidence-and-policy-implications/) restated these numbers in a June 2026 policy brief as the best experimental estimate available.

## Your evening decides when you stop more than the app does

A one-week field study of 46 people asked why they ended social media sessions. Most reasons had to do with their wider situation rather than anything inside the app ([Rixen et al., ACM](https://dl.acm.org/doi/10.1145/3604275)). The authors argue that tools to cut scrolling should account for the person's circumstances, too.

A CHI 2025 study tested this with 72 people over seven days, using an Android app that showed a "take a break" overlay after 15 minutes of continuous scrolling ([Scrolling in the Deep, ACM](https://dl.acm.org/doi/10.1145/3706598.3713187)). People who were at home and in a low mood took longer to stop after the prompt. If low mood is part of your pattern, [what 44 studies found about doomscrolling and anxiety](https://ivehadit.app/blog/does-doomscrolling-cause-anxiety-44-study-review) is worth a look.

### Sleepy people accepted break prompts more readily

If you can't stop scrolling at night, here's a mildly hopeful finding from the same study. Sleepy participants pushed back less against the prompt and were more accepting of it. A bedtime nudge has a better shot than you might expect. Both studies are small and short, so treat them as useful hints and keep the conclusions modest.

## The dopamine slot machine is an analogy, and a loose one

Many popular pages say every post triggers a burst of dopamine and that feeds rewire your brain. The slot-machine comparison is a metaphor nobody measured in the studies above. Claims that each post rewires your reward system go beyond what the human evidence shows. Even the Commission chooses its words with care, saying the design "may lead to" compulsive behaviour.

So is infinite scroll addictive? The defensible answer is that it's built to remove stopping points, and the behaviour it produces looks habit-driven and hard to control. That's enough to explain your evenings without any brain-chemistry drama.

### Is scrolling an addiction? Two markers say it behaves like one

The Allcott team looked for two signs of addiction: habit formation and self-control problems. They found both. Cutting back also reduced self-reported addiction-type behaviour (using longer than intended, trouble putting the phone down, using it to fall asleep, mindless use) and slightly improved overall wellbeing.

In March 2026 a Los Angeles jury found Meta and YouTube negligent in the design of their platforms, in a case brought by a young woman who said she became addicted as a small child ([NPR](https://www.npr.org/2026/03/25/nx-s1-5746125/meta-youtube-social-media-trial-verdict)). Damages were $6 million, with Meta held 70% responsible. That's one plaintiff and one jury, so read it as legal context rather than scientific proof ([The Conversation](https://theconversation.com/landmark-lawsuit-finds-that-social-media-addiction-is-a-feature-not-a-bug-279390)).

## How to stop scrolling, starting with what has the best evidence

The strongest tool in the research is pleasingly dull. In the Allcott experiment, being offered a tool to set your own daily app limits cut use by 22 minutes a day (16%), and the effect was still 19 minutes in the final week. 78% set limits that actually bit and kept using them, and people were willing to pay to keep the tool.

1. Pick tonight's limit at lunchtime: decide when you're calm, because pre-set limits are the best-tested fix we have.
2. Make the limit annoying to break: the EU found TikTok's own tools are easy to dismiss and add little friction, so use a passcode someone else sets or a blocker.
3. Remove the endless part: regulators propose switching off infinite scroll and autoplay, and you can do the same by blocking feeds while keeping messages, search and links.
4. Check the real numbers weekly: people underestimate their future use, so look at your screen-time report instead of guessing.
5. Give yourself an ending: the feed won't supply one, so set a timer, plan the next thing, or charge the phone outside the bedroom.
6. Expect the first weeks to be the hardest: lower use carried over for weeks after the incentive ended, so the effort compounds.

> **Plan for your risky evenings** Evenings at home when you're feeling low are when break prompts worked worst in the CHI 2025 study. Use your strictest block for those nights. And since sleepy people took prompts more kindly, switch on bedtime reminders even if they've felt pointless before.

If you're wondering whether the scrolling is doing you harm beyond the lost time, I wrote up [what two recent studies found about passive scrolling](https://ivehadit.app/blog/is-passive-scrolling-bad-for-you).

- Feeds are designed without stopping points, and the EU's preliminary findings say this shifts people into autopilot.
- Scrolling is a habit that strengthens with use, which also means cutting back today makes tomorrow easier.
- Self-control problems account for about a third of average social media use, with huge differences between people.
- Dopamine and slot-machine explanations are analogies, so be wary of pages that present them as measured facts.
- Pre-set limits that are hard to dismiss have the best evidence, followed by removing the endless feed and building your own stopping cues.

If you'd like the feed gone while the rest of the internet stays open, that's what I built I've had it to do. It closes social media feeds and leaves everything else alone. ([How it works](https://ivehadit.app/start))

## Questions people ask

### Why can't I stop scrolling even when I'm not enjoying it?

Scrolling runs largely on habit, and habits start before you consciously decide anything. The feed also never offers an ending, so you tend to stop only when something in your surroundings interrupts you.

### Why do I keep scrolling at night?

A 2025 field study found people at home and in a low mood took longer to stop after a break prompt. The same study found sleepy people accepted break prompts more readily, so bedtime reminders are worth switching on.

### Is scrolling an addiction?

In the largest experiment, researchers found both markers they used for addiction, habit formation and self-control problems. Self-control problems explained about 31% of average use, but the amount varied hugely between people.

### Do screen-time limits actually work?

Pre-set limits cut social media use by 22 minutes a day (16%) in a 12-week experiment, and most people kept using them. Limits that are easy to tap past work poorly, according to the EU's findings on TikTok.

### How long does it take to break the scrolling habit?

No study gives an exact timeline. In the Allcott experiment, reduced use carried over for at least six weeks after payments to cut back ended, which suggests the first weeks are the hardest and the effort compounds.

## Sources

1. [Commission preliminarily finds TikTok's addictive design in breach of the Digital Services Act](https://digital-strategy.ec.europa.eu/en/news/commission-preliminarily-finds-tiktoks-addictive-design-breach-digital-services-act), European Commission, 2026-02-06
2. [Commission preliminarily finds the addictive design of Instagram and Facebook in breach of the Digital Services Act](https://digital-strategy.ec.europa.eu/en/news/commission-preliminarily-finds-addictive-design-instagram-and-facebook-breach-digital-services-act), European Commission, 2026-07-10
3. [Digital Addiction (American Economic Review 112(7), working-paper version)](https://web.stanford.edu/~gentzkow/research/DigitalAddiction.pdf), Allcott, Gentzkow and Song, American Economic Review, 2022-07
4. [Digital addiction: Evidence and policy implications](https://www.hamiltonproject.org/publication/paper/digital-addiction-evidence-and-policy-implications/), The Hamilton Project (Brookings), 2026-06
5. [The Loop and Reasons to Break It: Investigating Infinite Scrolling Behaviour in Social Media Applications and Reasons to Stop](https://dl.acm.org/doi/10.1145/3604275), Proceedings of the ACM on Human-Computer Interaction (MobileHCI), 2023
6. [Scrolling in the Deep: Analysing Contextual Influences on Intervention Effectiveness during Infinite Scrolling on Social Media](https://dl.acm.org/doi/10.1145/3706598.3713187), CHI '25, ACM, 2025
7. [Jury finds Meta and Google negligent in social media harms trial](https://www.npr.org/2026/03/25/nx-s1-5746125/meta-youtube-social-media-trial-verdict), NPR, 2026-03-25
8. [Landmark lawsuit finds that social media addiction is a feature, not a bug](https://theconversation.com/landmark-lawsuit-finds-that-social-media-addiction-is-a-feature-not-a-bug-279390), The Conversation, 2026-03-26